16:04:23 <quidam> #startmeeting 13/Sep/2019
16:04:23 <Trisnity> Meeting started Fri Sep 13 16:04:23 2019 UTC.  The chair is quidam. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:04:23 <Trisnity> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:04:27 <quidam> hi everybody!
16:04:56 <Ark74> hello!
16:06:49 <valessio> o/
16:10:06 <quidam> valessio and I are in the FSF office working on the trisquel website
16:11:52 <dctrud> Hi, happy Friday all
16:12:41 <valessio> preview:  http://wwwdev.trisquel.info  - Suggestions, criticisms and patches are welcome at:   https://devel.trisquel.info/artwork/trisquel-d8-theme/issues
16:13:20 <daroal> hi all!
16:14:54 <daroal> there's a lot of good work there, valessio!
16:15:54 <chaosmonk> valessio: looks nice! i'll add some comments on gitlab later
16:16:37 <chaosmonk> very noscript friendly
16:17:07 <chaosmonk> is it all css?
16:17:22 <valessio> yes
16:17:50 <quidam> we are having trouble with the fact that the current documentation is written in markup (mediawiki format)
16:17:56 <quidam> and drupal 8 seems to not support that?
16:18:21 <quidam> it is quite odd
16:19:15 <chaosmonk> quidam: what does it support? just html?
16:19:29 <dctrud> what does drupal want these days? Is it a case of having to pipe all the existing through pandoc? :-/
16:23:38 <quidam> it can support different inputs using modules and libraries
16:23:46 <quidam> and there is one for markdown, and other formats
16:23:51 <quidam> but I can't find the markup one
16:24:55 <Ark74> valessio, issue filled: https://devel.trisquel.info/artwork/trisquel-d8-theme/issues/10
16:26:12 <dctrud> quidam: chaosmonk has been driving forward a lot of package-helpers again this week
16:26:25 <dctrud> chasmonk: many thanks for your testing of the fewer things I did
16:26:32 <daroal> quidam: I remember switching formats quite some years ago because of something, but I don't recall the something
16:26:45 <daroal> wiki formats, I mean
16:27:10 <daroal> I also remember looking into pandoc batch processing for this same reason
16:29:23 <daroal> I don't think converting the existing docs would be a problem, but we should learn/remember why we chose that one in the first place just in case the reasoning stands
16:40:20 <dctrud> Oh, wanted to mention this might be my last hello here until November. Have family from the UK visiting for a month, and will be hanging out with them at lunch when they are here, away from the screen.
16:41:41 <chaosmonk> dctrud: have a good time! thanks for all your work so far, especially getting the kernel working
16:43:20 <daroal> dctrud: enjoy the extra family time :)
16:43:45 <dctrud> chasmonk: thank you, and for your work too... I had a need for a newer yesterday gpg2 to test some code on a project and was thinking about looking at that, and then there it was already :-)
16:44:19 <chaosmonk> quidam: after dctrud's MRs, all of the prereqs for the netinstaller <https://devel.trisquel.info/trisquel/package-helpers/blob/flidas/README.md#Netinstall> build for me. if you push those to the repo, i can start working on make-debian-installer
16:44:53 <dctrud> awesome - that's an exciting step to reach
16:46:50 <chaosmonk> the installer itself probably won't work with out a few more packages. i assume we need grub2, which i can build with dpkg-buildpackage, but which fails with sbuild
16:47:16 <dctrud> oh - I think I managed an sbuild of grub2
16:47:58 <dctrud> but I think that was by also modifying the sbuild environment to point at my sort of personal assembly of trisquel....https://www.randomroad.net/trisquel_test/
16:48:28 <chaosmonk> dctrud. last night when i tried with sbuild i got "FAIL: grub_cmd_set_date"
16:48:34 <dctrud> I have been running essentially the ubuntu mirror + purge, than adding the current state of `package-helpers` and `trisquel-packages` to that
16:48:54 <dctrud> hmmm... let me check my logs a second
16:50:15 <dctrud> hmm - I do have debs in my `~/Git/package-helpers/helpers/PACKAGES/grub2` ... but not sure what makes that work for me, not you
16:51:41 <chaosmonk> hm. since the build appeared to fail i just went to bed without checking to see if debs were produced. booting up my etiona machine to check
16:52:37 <dctrud> ohhh... `FAIL: grub_cmd_set_date` is a test suite failure
16:52:48 <dctrud> I have `PASS: grub_cmd_set_date`
16:53:50 <dctrud> if your etiona machine is a VM.... is it failing trying to update the VM's hardware clock date or something?
16:54:04 <chaosmonk> it's not a vm
16:54:32 <dctrud> see also: https://lists.gnu.org/archive/html/bug-guix/2017-06/msg00278.html
16:54:51 <dctrud> looks like Guix disabled that test
16:55:58 <dctrud> debian are also skipping it: https://salsa.debian.org/grub-team/grub/commit/2e487ee0ff9d56322b4fae3252df0ad92422507d
16:56:14 <dctrud> guess worth skipping it via the Trisquel package helper?
16:56:53 <chaosmonk> yeah, sounds like it
16:58:26 <dctrud> have a good weekend all - back to the day job and into a meeting. Cheers.
16:59:07 <chaosmonk> thanks dctrud
17:02:39 <Ark74> o/
17:04:16 <daroal> bye!
18:04:11 <chaosmonk> quidam: are you around?
18:13:05 <quidam> chaosmonk: sorry, having a weird day. im at lunch
18:16:04 <chaosmonk> quidam: no problem, sorry to hear that. i was going to ask if we could go over this weeks MRs, but there's no rush. have a good lunch
19:30:25 <quidam> chaosmonk: I can look now
19:31:18 <chaosmonk> quidam: thanks! there are a couple for trisquel-packages, and more for package-helpers
19:54:27 <chaosmonk> quidam: removed the other changelog line for grub2 (and software properties, where i also had two changelog lines)
19:54:42 <quidam> cool
19:54:54 <quidam> with the base-data, there is an important change to make there
19:55:32 <chaosmonk> is that part of grub?
19:55:35 <quidam> the current partman recipe is wrong, it causes / to not be encrypted when using full disk encryption
19:55:46 <quidam> I think it would be enough to have a separate /boot there
19:56:05 <quidam> and then / can also be xfs
19:56:39 <quidam> and / should then have $lvmok{ }
19:58:47 <chaosmonk> quidam: sorry, that's over my head
19:59:17 <quidam> I added a comment with my suggestion
19:59:44 <quidam> brb
20:00:36 <chaosmonk> quidam: where should the comment be? on the grub2 MR i only see you comment about the changelog line
20:01:25 <quidam> on the file being commented 9.0/trisquel-base-data/data/lib/partman/recipes/20trisquel
20:01:28 <quidam> chaosmonk:
20:03:59 <chaosmonk> quidam: oh, i see. that one is dctrud's. i don't think he's around rn
20:04:08 <quidam> oh, sorry
20:04:21 <quidam> I misread
20:04:52 <quidam> the gnupg2 one... I don't think that change is really worth maintaining
20:04:56 <quidam> we could just drop the helper
20:05:42 <quidam> then again it doesn't harm
20:05:47 <quidam> meh
20:05:52 <quidam> merging
20:06:45 <bandali> hey quidam, any updates with all things icecat-related?
20:07:07 <quidam> bandali: no, sorry. trying to find time for everything
20:07:21 <bandali> ah please don’t be sorry, totally understand :)
20:07:30 <bandali> thanks for all you’re already doing
20:07:43 <chaosmonk> re update-manager, prevent-autoconnection.patch still gets applied later in the script. remove-livepatch.patch is a new patch to get rid of ubuntu-specific livepatch integration which was also causing some of the other patches to fail
20:08:14 <quidam> chaosmonk: but the patch file is not in the mr, is it?
20:08:34 <chaosmonk> crap, did i forget to commit the patch? sorry, one sec
20:10:18 <chaosmonk> quidam: sorry about that. fixed
20:40:26 <chaosmonk> quidam: for packages that don't have a helper and don't depend on any packages that have a helper, should those all be building automatically, or is there a reason some might not be in the repo?
20:42:57 <quidam> if a package doesn't have a helper it should be fetched directly from the ubuntu repo. if it is not the case it is usually that there was a helper and that put the package in the blacklist
20:43:08 <quidam> so it needs manual resolution
20:43:49 <chaosmonk> makes sense
20:46:22 <chaosmonk> quidam: not sure if this got buried earlier, but the prerequisites for the netinstaller listed in the package-helpers README all build now, with dtcrup's recent MR for rpl. does that mean that if those packages are pushed to the repo, all that's left to do in order to create a netinstall image is get make-debian-installer working?
20:50:56 <quidam> chaosmonk: that would be the case, yes
20:51:06 <quidam> they need to be run in order, though
20:51:26 <quidam> iirc, first a couple libraries, then the netinst packages, then debian-installer
20:51:47 <quidam> debian-installer produces a tarball separately from the usual deb files
20:52:43 <chaosmonk> quidam: okay, can you push those today then so that i can try that this weekend?
20:55:04 <chaosmonk> quidam: oh, i might have misunderstood you
20:56:27 <chaosmonk> which steps can i do locally, and which steps can only be done by jenkins?
21:15:43 <quidam> chaosmonk: sorry, I'm getting distracted
21:16:17 <quidam> which MRs need to be merged for netinst?
21:18:12 <chaosmonk> quidam: just the rpl one, you already merged it https://devel.trisquel.info/trisquel/package-helpers/merge_requests/247
21:18:38 <quidam> chaosmonk: then the next fixes are on the script that launches jenkins builds
21:18:39 <chaosmonk> that one was blocking make-netcfg
21:18:47 <quidam> it seems to be broken
21:19:16 <chaosmonk> is that why packages with helpers haven't been getting pushed to the repo?
21:21:39 <chaosmonk> quidam: is that script somewhere in gitlab?
21:23:23 <quidam> it may not be, I'm not sure
21:23:29 <quidam> one sec
21:23:44 <quidam> no, it is not
21:24:14 <quidam> I can publish it, but it is such a piece of crap that I wonder if it is better to start from scratch
21:26:06 <quidam> https://devel.trisquel.info/trisquel/trisquel-builder/blob/master/watchdog.py
21:26:36 <quidam> beware of that thing
21:27:07 <chaosmonk> quidam: it does look like nvidiafb is free. it sound like dctrud might not be active for a while, so maybe you should merge his MR as is and then remove that line after, assuming you aren't able to edit his MR
21:30:10 <quidam> merged
21:31:49 <chaosmonk> quidam: does the watchdog script need to be fixed/rewritten before creating netinstall images, or can the necessary scripts be triggered manually?
21:32:01 <quidam> they can be launched manually
21:32:15 <quidam> chaosmonk: where is prevent-autoconnect.patch used?
21:33:39 <quidam> never mind, found
21:33:45 <chaosmonk> https://devel.trisquel.info/chaosmonk/package-helpers/blob/fix-update-utilites/helpers/make-update-manager#L37
21:33:55 <chaosmonk> k
21:33:56 <quidam> the viewer collapsing pieces of code doesn't help sometimes
21:34:00 <chaosmonk> yeah
21:35:12 <chaosmonk> quidam: can i just run all of those scripts in order on my machine to create a netinstall image, or do you need to push certain things to the repo?
21:36:46 <chaosmonk> once i have a netinstall image, i can see if/where the actual installation fails and work on that this weekend
21:37:24 <quidam> chaosmonk: you can run them all, but to create a netinst they need to be in the repos
21:37:37 <quidam> you can set up a local repo, but it is quite a hassle
21:39:54 <chaosmonk> ok, is there any way you could run them and push them to the repo today? i know it's getting a little late in the afternoon on your coast
21:41:37 <quidam> sure, still going through MRs
21:41:53 <chaosmonk> awesome, thanks
21:51:28 <quidam> building locally first...
21:52:27 <quidam> chaosmonk: do you want to merge the mate desktop one and then send further corrections, or wait?
21:53:04 <chaosmonk> wait on that one, still needs work
21:54:18 <quidam> building on jenkins now
21:55:14 <dctrud> Hi - I tried to make a netinstall against all the MRs, as I have them in a repo... it needed something from ubiquity that wasn't building.... but I'm afraid I forget what.
21:56:18 <dctrud> Please go ahead and edit / rebranch any of the MRs that needs changes, as chaosmonk said, I'm not going to be at the keyboard much other than for work for a while.
21:57:26 <quidam> dctrud: the only pending one is the trivial additions one
21:57:28 <quidam> I can edit it
21:57:40 <chaosmonk> dctrud: yeah, make-ubiquity doesn't work yet. i thought it would only be needed for the graphical installer, but you're saying the blocks netinstaller also?
22:00:01 <quidam> it should not block the netinst
22:00:55 <dctrud> The error trying to sbuild debian-installer is `Unable to locate package oem-config-check`
22:01:05 <dctrud> oem-config-check comes from ubiquity
22:01:13 <dctrud> https://packages.ubuntu.com/bionic/oem-config-check
22:02:08 <quidam> oh... welp
22:02:54 <quidam> I merged as far as I could for the day, and all the netinst packages are being built by jenkins
22:03:17 <quidam> for a test, you should be able to remove the oem-config-check requirement and build debian-installer with sbuild
22:03:24 <quidam> as soon as the builds finish
22:03:46 <quidam> but for now I have to go
22:04:27 <chaosmonk> thanks quidam
22:04:30 <quidam> I cannot tell you how happy I am to see so much progress being done! thank you so much!
22:06:49 <quidam> #endmeeting